Observing the X- and Gamma-Ray Sky Diffuse emission J. Knödlseder Centre d’Etude Spatiale des Rayonnements Toulouse (France)
Observing the X- and Gamma-Ray Sky
Diffuse emission
J. KnödlsederCentre d’Etude Spatiale des Rayonnements
Toulouse (France)
Lecture Outline
I. What is diffuse emission ?
II. Diffuse emission processes
III. The X- and Gamma-Ray Sky
• Sky images
• The galactic emission spectrum
IV. The nature of galactic X- and Gamma-Ray emissions
• The Galactic Ridge X-ray emission (GRXE)
• The hard X-ray Sky
• Positron annihilation
(imaging diffuse emission)
• Galactic Radioactivities
• The MeV - GeV Sky
• The TeV Sky
V. Summary & Bibliography
Diffuse or not diffuse - that is the question
Allsky image in visible light (Mellinger 2000)
A working definition for diffuse emission
Dictionary:
Diffuse = widely spread; not localized or confined; with no distinct margin
Astrophysicist:
“all emission processes that are related to interstellar (-planetary, -galactic) matter”• emission of gaz and dust (thermal, non-thermal)
• emission related to magnetic fields (synchroton)
• emission related to diffuse stellar ejecta (particle diffusion)
• also applicable to extragalactic diffuse (e.g., intergalactic matter in clusters)
• also applicable for cosmic backgrounds (e.g., primordial matter for CMB)
Astronomer:
“all emission that I cannot resolve into individual (point-) sources”• depends on instrument characteristics (angular resolution, sensitivity)
• is not of much help for an astrophysicist
Diffuse emission processes
Interaction of high-energy CR electrons andnucleons with gas and radiation in the ISM:
Inverse Comptonelectron scattering
electronBremsstrahlung
Pion (!0) productionand decay p + p " p + p + !0 " 2# Ep > 300 MeV
Continuum emission Line emission
Excitation of electrons and nucleons in anatom; antimatter annihilation:
ionic lines(below 10 keV)
nuclear radioactivedecay
nuclear excitation
positron-electronannihilation(511 keV line)
A high-energy gallery of the sky
keV
TeV
The Soft X-Ray Sky (1 - 3 keV)ROSAT
The X-Ray Sky (2 - 20 keV)HEAO-1
The Hard X-Ray Sky (25 - 50 keV)SPI / INTEGRAL (2 yr)
The Hard X-Ray Sky (50 - 100 keV)SPI / INTEGRAL (2 yr)
The Hard X-Ray Sky (100 - 200 keV)SPI / INTEGRAL (2 yr)
The Hard X-Ray Sky (200 - 300 keV)SPI / INTEGRAL (2 yr)
The Gamma-Ray Sky (300 - 400 keV)SPI / INTEGRAL (2 yr)
The Gamma-Ray Sky (400 - 500 keV)SPI / INTEGRAL (2 yr)
The Gamma-Ray Sky (511 keV line)SPI / INTEGRAL (2 yr)
The Gamma-Ray Sky (514 - 1000 keV)SPI / INTEGRAL (2 yr)
The Gamma-Ray Sky (1 - 3 MeV)COMPTEL / CGRO (6 yr)
The Gamma-Ray Sky (1809 keV line)COMPTEL / CGRO (9 yr)
The Gamma-Ray Sky (3 - 10 MeV)COMPTEL / CGRO (6 yr)
The Gamma-Ray Sky (10 - 30 MeV)COMPTEL / CGRO (6 yr)
The HE Gamma-Ray Sky (30 - 100 MeV)EGRET / CGRO (4 yr)
The HE Gamma-Ray Sky (100 - 300 MeV)EGRET / CGRO (4 yr)
The HE Gamma-Ray Sky (300 - 1000 MeV)EGRET / CGRO (4 yr)
The HE Gamma-Ray Sky (1 - 30 GeV)EGRET / CGRO (4 yr)
The VHE Gamma-Ray Sky (0.1 - 20 TeV)H.E.S.S.
The galactic diffuse emission spectrum
1038 erg s-1 1038 erg s-1
X-ray galactic ridge emission
X-ray (2-10 keV) emission components• point sources (X-ray binaries)• unresolved (or diffuse) emission
Galactic ridge X-ray emission (GRXE)• exponential disk & bulge components• confined to the inner disk (|l| < 60°)• disk scale height z0 ~ 100 - 300 pc• luminosity ~ 1038 erg s-1 (2 - 10 keV) (few % of resolved sources luminosity)
Origin of GRXE• unresolved point sources?• truly diffuse emission?
HEAO-1 2-50 keV map (Allen et al. 1994)
EXOSAT 2-6 keV map (Warwick et al. 1985)
Deep X-ray surveys (XMM & Chandra)
l=28.5° (Ebisawa et al. 2005)
• XMM & Chandra detect new faint point sources and prominent diffuse emission
• Only 10-20% of flux originates from point sources, 80-90% of the emission is diffuse
• Soft (< 2 keV) point sources are of galactic origin
• Hard (2-10 keV) point sources are of extragalactic origin
• Prominent emission lines from highly ionized heavy elements
GC (Muno et al. 2004)
GC
l=28.5°
Point-source origin
Point source hypothesis
• Candidate must have a thin thermal plasma spectrum with iron line emission
• Candidate population requires rapid steepening of log N-log S at low flux (<3 x 10-15 erg s-1 cm-2)
(Ebisawa et al. 2005) (Koyama et al. 1986)
1000 sources
Candidats
• NS binaries (1036-38 erg s-1): rarely show iron line, most of them are individually resolved
• RS CVn / CVs (1030-32 erg s-1): resolved by Chandra/XMM, but not numerous enough
• Low luminosity population (<1030 erg s-1): >109 sources required within Galaxy
Diffuse origin
Inverse Compton scattering of microwave background, FIR photons, starlight
• fall short by 2 orders of magnitudes
• CR scale-height of > 1 kpc does not match the GRXE scale-height
Synchrotron radiation
• requires ~ 1014 eV electrons $ unclear whether they exist (solar modulation)
• large energy input required to sustain electron population $ ionisation of ISM
Thermal equilibrium plasma
• requires T ~ 107 - 108 K $ plasma exceeds escape velocity from galactic plane
• requires P/k ~ 105 cm-3 K $ exceeds pressure of other ISM components
• required energy density ~ 10 eV cm-3 $ 1-2 orders of magnitudes higher than CR, B, n
CR interactions with interstellar medium
• interactions of low-energy CR e-, in-situ accelerated e-, or heavy ions with ISM
• hard X-ray emitting SNR AX J1843.8-0352: possible link between GRXE and SNRs
Finally point sources?
RXTE/PCA 3-20 keV image (Revnivtsev et al. 2006)
Morphology• tri-axial bar/bulge & exponential disk• distribution very similar to NIR (e.g., COBE 3.5 !m)• bar tilt angle: 29° ± 6° (COBE NIR data: 20° ± 10°)• exponential disk scale-height: z = 130 ± 20 pc• position of Sun above gal. Plane: z0 = 20 ± 7 pc
Luminosities• LX,bulge = (3.9 ± 0.5) x 1037 erg s-1
• Mbulge = (1.0 - 1.3) x 1010 M!
• LX/M! = (3.5 ± 0.5) x 1027 erg s-1
• Comparable with cumulative emissivity per unit stellar mass of point X-ray sources in solar neighbourhood (coronally active late-type binaires and CVs)
spectra divided by 2
INTEGRAL resolves the hard X-ray ridge
RXTE & OSSE (Valinia & Tatischeff 2001)
Hard X-ray emission components
• < 10 keV RS (= GXRE)
• 10 - 200 keV PL (exponentially cut-off powerlaw)
• > 200 keV HE (high-energy flattening)
IBIS (Lebrun et al. 2004)
PL (exponentially cut-off powerlaw)
• IBIS detects many point source towards the galactic bulge region
• Most of the total emission is attributed to point sources 20 - 40 keV: 87% attributed to point sources
• By combining IBIS (point sources) and SPI (total emission) 100 - 200 keV: 86% attributed to point sources
The hard X-ray to soft #-ray transition
OSSE spectra (Kinzer et al. 1999)
SPI 200-300 keV SPI 300-400 keV
Emission components
• < 200 keV (hard X-rays) PL (exponentially cut-off powerlaw)
• 200 - 511 keV (soft #-rays) Pscont (Positronium continuum, towards bulge only)
• > 511 keV (#-rays) HE (high-energy power-law tail)
Is the transition hard X-ray $ soft #-raya point source $ diffuse emission transition?
Antimatter annihilation in the Milky-Way
e-
#
#
e+
e-
e+
ortho-positronium
3/4
para-positronium
1/4
e-
e+
#
# #
#
#
• Direct annihilation
• Annihilation via positronium (Ps) formation
Positronium continuum
E < 511 keV
Annihilation line
E = 511 keV
Positron annihilation: spatial distribution
Observations
• No point sources seen (SPI & IBIS)
• Continuum and line are spatially consistent
• Galactic bulge dominates emission
• Only small signal from galactic disk (~3%)
• B/D luminosity ~ 3 - 9
Implications
• Positron annihilation distribution is unique Once we identify the source we certainly learn something new! (new population, new mecanism, new physics, …)
• Weak galactic disk signal compatible with 26Al decay
SPI 511 keV image (Knödlseder et al. 2005)SPI Pscont image (Weidenspointner et al. 2006)
Indirect imaging: deconvolving SPI data
Iteration 1 Exposure
Log likelihoodFlux
Indirect imaging: deconvolving SPI data
ExposureIteration 5
Flux Log likelihood
Indirect imaging: deconvolving SPI data
ExposureIteration 10
Flux Log likelihood
Indirect imaging: deconvolving SPI data
ExposureIteration 17
Flux Log likelihood
Indirect imaging: deconvolving SPI data
ExposureIteration 25
Flux Log likelihood
Indirect imaging: deconvolving SPI data
ExposureIteration 40
Flux Log likelihood
Indirect imaging: deconvolving SPI data
ExposureIteration 70
Flux Log likelihood
Indirect imaging: deconvolving SPI data
ExposureIteration 100
Flux Log likelihood
Suppressing noise
Richardson-Lucy (iteration 17) MREM (Knödlseder et al. 1999)
Model fitting (bulge + old disk) Model fitting (halo + old disk)
Positron annihilation: spatial distribution
Observations
• No point sources seen (SPI & IBIS)
• Continuum and line are spatially consistent
• Galactic bulge dominates emission
• Only small signal from galactic disk (~3%)
• B/D luminosity ~ 3 - 9
Implications
• Positron annihilation distribution is unique Once we identify the source we certainly learn something new! (new population, new mecanism, new physics, …)
• Weak galactic disk signal compatible with 26Al decay
SPI 511 keV image (Knödlseder et al. 2005)SPI Pscont image (Weidenspointner et al. 2006)
Positron annihilation: spectral distribution
SPI spectrum (Jean et al. 2006)
SPI spectral fitting
• Energy 510.98 ± 0.03 keV
• FWHMn 1.3 ± 0.4 keV
• FWHMb 5.4 ± 1.2 keV
• Fluxn 7.2 x 10-4 ph cm-2 s-1
• Fluxb 3.5 x 10-4 ph cm-2 s-1
Interpretation
• Narrow line (1.1 keV) - thermalised positrons - consistent with 8000 K warm ISM (neutral & ionised)
• Broad line (5.1 keV) - inflight positronium formation - consistent with 8000 K warm ISM (only neutral, quenched if gaz is fully ionised)
• Narrow / broad line fraction ~ 2 consistent with 8000 K warm ISM (50% ionised)
Radioactive decay in the Milky-Way
Distribution of 26Al and 60Fe in the ISM
• velocity of 1 km/s corresponds to a distance of 1 pc with 1 Myr
• SN ejection velocities: 1000 - 10000 km s-1 (but slow down)
• WR wind velocities: several 1000 km s-1
• SN or wind blown bubbles: 10 - 100 pc
• 26Al and 60Fe should lead to diffuse emission, nuclei probably thermalised
• Short livetime isotopes (<100 yr, such as 44Ti, 7Be, 22Na, 56,57Co): point-like emission (<pc)
60Fe decay scheme26Al decay scheme
26Al decay 1809 keV line emission
1809 keV line: radioactive 26Al production
• H and C-burning nucleosynthesis
• Hydrodynamic and explosive
• Stellar wind ejection (O, LBV, WR)
• Supernovae ejection (type II, Ib/c)
• Probe stellar mixing processes
• Traces massive stars
COMPTEL image (Knödlseder et al. 1999)
DMR microwave image (Bennett et al. 1992)
26Al production and massive stars
• 1809 keV emission correlates to microwave free-free emission
• Free-free: ionised ISM (O stars, M>20 M!)
• Y26 = 10-4 M! /O7V
Calibrating stellar models
1809 keV #-rays (COMPTEL)
Understanding 26Al nucleosynthesis in Cygnus
• Bright 1809 keV line feature
• Massive star population of Cygnus region is known (IR surveys)
• Estimate expected 1809 keV line flux using nucleosynthesis models and stellar population models (Cerviño et al. 2000; Knödlseder et al. 2002)
• Validate model using multi-wavelength properties (e.g. ionizing flux)
• 1809 keV flux underestimated by at least a factor of 2 (mixing?, stellar rotation?)
Infrared (IRAS) Radio (DRAO)
1809 keV line emission traces galactic rotation
INTEGRAL spectra (Diehl et al. 2006)
26Al kinematics
• Galactic rotation (v ~ 200 km s-1) leads to Dopples shifts (~ 1 keV)
• Expected average line shifts ± 0.3 keV (from CO)
• Measured line shifts ± 0.3 keV (SPI/INTEGRAL)
• Confirmation of galaxy-wide 26Al production (2.8 ± 0.8 M!)
• Using yield estimates (theory) this converts into SFR of 4 M! yr-1
60Fe: A long way to a faint radioactivity
SPI spectrum (Harris et al. 2005)
SPI/INTEGRAL and RHESSI measurements
• 60Fe / 26Al flux ratio ~ 10%
• 60Fe / 26Al abundance ratio ~ 0.23
Interpretation
• 60Fe only produced in core-collapse evens
• 26Al produced in core-collapse and WR winds
• Expected core-collapse 60Fe / 26Al ratio too large
• WR winds contribute significantly to galactic 26Al nucleosynthesis
(Prantzos 2004)
Diffuse MeV and GeV Gamma-Ray emission
Point sources
• Pulsars
• Supernova remnants
• AGN (extragalactic)
• unidentified sources
Diffuse emission processes
• inverse Compton
• Bremsstrahlung
• nuclear interactions lines
• !0 decay (> 300 MeV)
Spatial correlation between gaz and #-rays
Observations (EGRET):
• large scale spatial distribution well modelled by combination of ISM phases (assuming I & '2)
• fraction of unresolved point sources is small (unless distributed like the interstellar gas)
• spectrum does not vary (within relatively small uncertainties) in the Galaxy
• deviations from perfect fit
Implications:
• Gamma-Rays probe galactic CR and ISM distributions
• CR electron-to-proton ratio roughly constant throughout Galaxy
• assumption of dynamic balance (I & '2) between ISM and CR is reasonably correct (large matter density implies larger magnetic fields, allowing for larger CR energy density)
Spectral modelling: The conventional model
Electron spectrum
• E -1.6 : E < 10 GeV
• E -2.6 : E > 10 GeV
• agrees with locally measured spectrum
• satisfies synchrontron spectrum
Proton spectrum
• E -2.25
• agrees with locally measured spectrum
Model
• based on non #-ray data only
• fits only between 30 - 500 MeV
C model (Strong et al. 2000)
Spectral modelling: Hard CR spectrum model
Electron spectrum
• E -1.8 (harder w/r C-model above 10 GeV)
• differs from locally measured spectrum (high-energy e- undergo rapid E-loss)
• satisfies synchrontron spectrum (> 10 GeV spectrum unconstrained)
Proton spectrum
• E -1.8 : E < 20 GeV (harder w/r C-model)
• E -2.5 : E > 20 GeV
• agrees with locally measured spectrum (solar modulation allows for some freedom at low energies)
Model
• allow for harder e- and p specturm
• does not fit <30 MeV (& GeV bump)
HEMN model (Strong et al. 2000)
Spectral modelling: Steep low-energy e- model
Electron spectrum
• E -3.2 : E < 200 MeV (steeped w/r C-model)
• E -1.8 : E > 200 MeV (like HEMN model)
• differs from locally measured spectrum (high-energy e- undergo rapid E-loss)
• satisfies synchrontron spectrum (< 1 GeV spectrum unconstrained)
Proton spectrum
• E -2.25 (C-model)
• agrees with locally measured spectrum
Model
• allows for more low-energy e-
• ad hoc (no observational evidence)
• large power input into ISM (ionisation)
SE model (Strong et al. 2000)
A dark-matter scenario
Possible explanations of GeV excess
• different CR spectrum than local
• unresolved point-sources
• EGRET calibration error
• Dark Matter
Dark Matter Model
• WIMP annihilation: ( + ( " q + q " !0 " #
• WIMP mass 50 - 100 GeV best fits the EGRET data
• Derive WIMP distribution from #-ray distribution
• DM in halo and 2 elliptical rings (R = 4 & 14 kpc)
• DM distribution can explain rotation curve
(de Boer et al. 2005)
But … (Bergström et al. 2006)
• WIMP annihilation should also produce antiprotons
• Observed antiproton flux much too low w/r model
• Strange DM distribution (ressemblence to baryon distribution with bulge, thin and thick disk)
The first VHE survey of the Galaxy
H.E.S.S. image (Aharonian et al. 2005)
H.E.S.S. survey
• longitudes ±35°, latitiudes ±4°
• 10 sources from which 8 are new (all spatially resolved $ extended emission)
• clustering of sources along the galactic plane (young population)
• some plausible associations with SNRs and pulsars
VHE diffuse emissionH.E.S.S. discovery of diffuse emission
• Subtract point-like emission from sources
• Extended emission (in l and b) along gal. Plane
• Correlates with molecular gas (CS)
• Power law spectrum: ) = 2.3 ± 0.3
H.E.S.S. image (Aharonian et al. 2005)
Interpretation
• !0 decay following CR interaction with ISM
• Flux higher and harder than expected $ recent (~10,000 yr) CR acceleration at GC and diffusion
The nature of galactic X-/#-ray emission
faint point sources CR - ISM interactions???
SummaryHard X-ray emission - GRXE (E < 200 keV)• observationally, a diffuse (unresolved) component remains
• theoretically, diffuse emission is difficult to understand (pressure, gravitational binding)
• spatial distribution and spectrum consistent with population of weak X-ray point sources
Soft #-ray regime (200 keV < E < 511 keV)• diffuse positronium annihilation dominates (bulge region)
• still no e+ point sources detected (but diffusion make annihilation probably inherently diffuse process)
MeV domain (1 MeV < E < 30 MeV)• 26Al and 60Fe radioactive decays lead to diffuse line emission
• source of continuum emission unclear (unresolved MeV point sources?)
GeV domain (30 MeV < E < 30 GeV)• diffuse emission explained by CR interaction with ISM
• spectrum leaves room for additional components (Dark Matter?, point sources?)
TeV domain (E > 30 GeV)• individual point sources identified (SNRs, pulsars)
• diffuse emission component that correlates with molecular clouds
Bibliography (some selected articles)
• Galactic ridge X-ray emission Lebrun et al. 2004, Nature, 428, 293 Ebisawa et al. 2005, ApJ, 635, 214 Revnivtsev et al. 2006, A&A, in press (astro-ph/0510050)
• Soft gamma-ray emission Kinzer et al. 1999, ApJ, 515, 215
• Positron annihilation Knödlseder et al. 2005, A&A, 441, 513 Jean et al. 2006, A&A, 445, 579
• Galactic radioactivity Harris et al. 2005, A&A, 433, L49 Diehl et al. 2006, Nature, 439, 45
• MeV and GeV galactic diffuse emission Strong et al. 2000, ApJ, 537, 763 deBoer et al. 2005, A&A, 444, 51
• TeV galactic diffuse emission Aharonian et al. 2006, Nature, 439, 695